If you have image processing toolbox
M = rand(1000, 1000);
M_new = blockproc(M, [10 10], @(x) mean(x.data, 'all'))
Alternative Solution 1: (surprisingly the fastest)
M_new = conv2(M, ones(10)/100, 'valid');
M_new = M_new(1:10:end, 1:10:end);
Alternative Solution 2:
M_C = mat2cell(M, 10*ones(1,100), 10*ones(1,100));
M_new = cellfun(@(x) mean(x, 'all'), M_C);
Alternative Solution 3:
M_new = zeros(size(M)/10);
for i=1:100
for j=1:100
M_new(i, j) = mean(M(10*(i-1)+1:10*i,10*(j-1)+1:10*j), 'all');
end
end
